Standard
Tagged PDF
A Tagged PDF is a PDF document that carries an underlying logical structure tree — sometimes called a "tag tree" — describing the organisation and semantic meaning of the content in addition to its visual appearance. The concept and its technical rules are defined in the core PDF specification, ISO 32000 (Clause 14.8 governs tagged PDF, covering standard structure element types, structure attributes, and word/character disambiguation). A tagged PDF is therefore not a separate file format but a set of requirements layered onto an ordinary PDF.
The structure is expressed as a hierarchy of structure elements rooted at a structure tree root, referenced from the document catalog via the StructTreeRoot entry. Each element is identified by a structure type that conveys its role: headings, paragraphs, lists and list items, tables with their rows and cells, figures, links, and so on — conceptually similar to how HTML elements describe a web page. Tags do several essential jobs at once. They establish an explicit, author-intended reading order independent of how content happens to be positioned on the page; they distinguish real content from artifacts such as page numbers, decorative rules, or background graphics; they carry semantic roles so a heading is announced as a heading and a data table's row and column relationships are preserved; and they provide the attachment points for accessibility properties such as alternative text and language.
Tagging is the foundation on which document accessibility standards are built: PDF/UA (ISO 14289) requires a correctly Tagged PDF, and the WCAG PDF techniques rely on tagging to satisfy their success criteria. Its importance is clearest in its absence. Without tags, a screen reader can generally only extract the raw text from the page content streams. That raw text may not follow the intended reading order — multi-column layouts, sidebars, and captions can be jumbled — and it carries no structural meaning, so a user cannot tell a heading from body text, cannot navigate by headings, cannot understand which header cells relate to which data cells, and receives no indication that an image is present or what it conveys.
A properly Tagged PDF resolves all of these by encoding the document's logical structure directly into the file. Tagging alone does not guarantee accessibility — tags can be incomplete, incorrect, or in the wrong order — but a document cannot be robustly accessible without it. It is the necessary substrate that alt text, reading order, headings, and table semantics all depend upon.
How Normfluss checks for it
Normfluss first determines whether the document is tagged at all, then examines the structure tree for completeness and correctness — that content is enclosed in appropriate structure elements, that headings, lists, and tables use the right element types and nesting, that reading order follows the tag sequence, and that decorative content is marked as artifact rather than left untagged. These structural findings are reported as conformance within a stated scope; issues that need human judgement, such as whether the tagged reading order is logical for a complex layout, are flagged for manual review.